A new regularized subband ANC algorithm with online secondary-path modeling: Performance analysis and application to buildings

Abstract

The wideband active noise control (ANC) system has been widely used in buildings and other indoor environments, which usually suffers from high computational complexity and low convergence speed due to the long channel responses and colored inputs. To solve this problem, the subband adaptive filtering has been employed in ANC systems. In this paper, a new regularized multiband-structured subband adaptive filter (MSAF) algorithm for ANC with online secondary-path modeling (SPM) is proposed. Difference equations describing the mean and mean square convergence behaviors of the adaptive algorithms for the ANC system in nonstationary environments are derived and these equations at the steady state are solved to obtain the mean square errors (MSEs) for the ANC controller and the SP estimator. Based on the analyses, practical methods for selecting the step-sizes and regularization parameters are further developed. The resulting variable regularized and variable step-size (VR-VSS-) MSAF for the controller and VSS-LMS algorithm for the SP estimator are hence employed in the proposed subband ANC system. Computer simulations are conducted to verify the effectiveness of the theoretical analysis. A scenario based on the real room impulse response in buildings is further considered and the usefulness of the proposed algorithm in practice is discussed.